Face scoring model training method, face scoring method and related device

Abstract

The invention provides a face scoring model training method, a face scoring method and a related device. Generating a pseudo label according to the first image marked with the scoring label, and distributing the pseudo label to the second image not marked with the scoring label to obtain a third image; obtaining a first scoring result according to the first image and the third image; performing feature extraction according to the first scoring result and the first feature to obtain a second feature, and performing feature fusion according to the second feature to obtain a fused feature; obtaining a second scoring result according to the fusion features; extracting multi-scale features of the training face image, and obtaining a third scoring result according to the third features, the second scoring result and the multi-scale features; obtaining a trained face scoring model according to the first scoring result, the second scoring result and the third scoring result; weighted fusion is performed by integrating a plurality of scoring results, and the importance and reliability of different features are combined, so that the accuracy and reliability of the target scoring result are ensured.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of image recognition technology, and in particular to a training method for a face scoring model, a face scoring method, and related devices. Background Art

[0002] Face scoring technology is a technology that combines image recognition and deep learning to score the appearance of a face image. Face scoring models usually require a large amount of data for training to achieve good results. In practical applications, there is a problem of insufficient labeled data in the training image database. In addition, complex image features and long sequence data will also greatly reduce the accuracy and reliability of face scoring models. Summary of the invention

[0003] The following is a summary of the subject matter described in detail herein. This summary is not intended to limit the scope of the claims.

[0004] The purpose of the present application is to solve at least one of the technical problems existing in the related art to a certain extent. The embodiments of the present application provide a training method for a face scoring model, a face scoring method and related devices, which can improve the accuracy and reliability of the predicted scores.

[0005] An embodiment of the first aspect of the present application is a method for training a face scoring model, comprising:

[0006] Acquire a face image to be trained, and input the face image to be trained into a face scoring model, wherein the face image to be trained includes a first image marked with a scoring label and a second image not marked with a scoring label;

[0007] Generate a pseudo label according to the first image, and assign the pseudo label to the second image to obtain a third image marked with the pseudo label;

[0008] Extracting features from the first image and the third image to obtain a first feature, and obtaining a first scoring result based on the first feature;

[0009] Performing feature extraction based on the first scoring result and the first feature to obtain a second feature, and performing feature fusion based on the second feature to obtain a fusion feature;

[0010] Performing feature extraction according to the fused feature to obtain a third feature, and obtaining a second scoring result according to the third feature;

[0011] Extracting multi-scale features of the training face image, and obtaining a third scoring result according to the third feature, the second scoring result and the multi-scale features;

[0012] The parameters of the face scoring model are adjusted according to the first scoring result, the second scoring result and the third scoring result to obtain a trained face scoring model.

[0013] According to some embodiments of the first aspect of the present application, generating a pseudo label according to the first image includes:

[0014] Perform weakly supervised learning on the classifier of the face scoring model according to the first image to generate a pseudo label.

[0015] According to some embodiments of the first aspect of the present application, assigning the pseudo label to the second image to obtain a third image marked with the pseudo label includes:

[0016] Filtering the pseudo labels according to the confidences corresponding to the pseudo labels to obtain filtered pseudo labels;

[0017] The filtered pseudo-label is assigned to the second image to obtain a third image marked with the pseudo-label.

[0018] According to certain embodiments of the first aspect of the present application, extracting features from the first image and the third image to obtain a first feature, and obtaining a first scoring result based on the first feature, includes:

[0019] Performing semi-supervised learning on the classifier of the face scoring model according to the first image and the third image to extract a first feature;

[0020] Classification prediction is performed based on the first feature to obtain a first scoring result.

[0021] According to certain embodiments of the first aspect of the present application, extracting a second feature based on the first scoring result and the first feature, and fusing the second feature to obtain a fused feature include:

[0022] Performing convolution on the first scoring result and the first feature to extract a second feature;

[0023] Converting the second feature to a feature space of the same dimension;

[0024] The converted second feature is subjected to feature fusion to obtain a fused feature.

[0025] According to certain embodiments of the first aspect of the present application, extracting a third feature according to the fused feature, and obtaining a second scoring result according to the third feature, comprises:

[0026] Training a classifier of a face scoring model according to the fused features, mapping the fused features to a class activation map, and obtaining a third feature;

[0027] Global maximum pooling is performed according to the third feature to obtain a second scoring result.

[0028] According to certain embodiments of the first aspect of the present application, adjusting the parameters of the face scoring model according to the first scoring result, the second scoring result, and the third scoring result to obtain a trained face scoring model includes:

[0029] Obtain a target scoring result according to a first preset weight parameter, a second preset weight parameter, a third preset weight parameter, the first scoring result, the second scoring result, and the third scoring result;

[0030] Obtaining a loss function value according to the target scoring result;

[0031] The parameters of the face scoring model are adjusted according to the loss function value to obtain a trained face scoring model.

[0032] An embodiment of the second aspect of the present application is a face scoring method, comprising:

[0033] Get the face image to be rated;

[0034] Inputting the face image to be rated into a trained face rating model for rating, and obtaining a rating result;

[0035] The trained face scoring model is obtained by training according to the face scoring model training method according to any one of claims 1 to 7.

[0036] An embodiment of the third aspect of the present application is an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the training method of the face scoring model described in the embodiment of the first aspect of the present application and the face scoring method described in the embodiment of the second aspect of the present application are implemented.

[0037] An embodiment of the fourth aspect of the present application is a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the training method of the face scoring model described in the embodiment of the first aspect of the present application and the face scoring method described in the embodiment of the second aspect of the present application are implemented.

[0038] The above scheme has at least the following beneficial effects: generating a pseudo label according to a first image marked with a rating label, assigning the pseudo label to a second image not marked with a rating label, and obtaining a third image marked with a pseudo label; performing feature extraction according to the first image and the third image to obtain a first feature, and obtaining a first rating result according to the first feature; performing feature extraction according to the first rating result and the first feature to obtain a second feature, and performing feature fusion according to the second feature to obtain a fused feature; performing feature extraction according to the fused feature to obtain a third feature, and obtaining a second rating result according to the third feature; extracting multi-scale features of the training face image, and obtaining a third rating result according to the third feature, the second rating result and the multi-scale features; adjusting the parameters of the face rating model according to the first rating result, the second rating result and the third rating result to obtain a trained face rating model; and ensuring the accuracy and reliability of the target rating result by weighted fusion through comprehensive integration of multiple rating results and combining the importance and reliability of different features. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings are used to provide further understanding of the technical solution of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the technical solution of the present application and do not constitute a limitation on the technical solution of the present application.

[0040] Figure 1 is a step diagram of a face scoring method provided in an embodiment of the present application;

[0041] Figure 2 is a step diagram of a training method for a face scoring model provided in an embodiment of the present application;

[0042] Figure 3 is a sub-step diagram of generating a pseudo label according to a first image provided by an embodiment of the present application;

[0043] Figure 4 It is a sub-step diagram of assigning a pseudo label to a second image to obtain a third image marked with the pseudo label provided by an embodiment of the present application;

[0044] Figure 5 It is a sub-step diagram of step S400 provided in an embodiment of the present application;

[0045] Figure 6 It is a sub-step diagram of step S500 provided in an embodiment of the present application;

[0046] Figure 7 It is a sub-step diagram of step S600 provided in an embodiment of the present application;

[0047] Figure 8 It is a sub-step diagram of step S800 provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0049] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims or the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0050] The embodiments of the present application are further described below in conjunction with the accompanying drawings.

[0051] The embodiments of the present application provide a face scoring method.

[0052] Reference Figure 1 , a face scoring method, comprising the following steps:

[0053] Step S110, obtaining a face image to be rated;

[0054] Step S120, inputting the face image to be rated into the trained face rating model for rating, and obtaining a rating result.

[0055] The trained face scoring model is obtained by training according to the following face scoring model training method.

[0056] Reference Figure 2 , the training method of the face scoring model includes the following steps:

[0057] Step S200, obtaining a face image to be trained, and inputting the face image to be trained into a face scoring model;

[0058] Step S300, generating a pseudo label according to the first image, assigning the pseudo label to the second image, and obtaining a third image annotated with the pseudo label;

[0059] Step S400, extracting features from the first image and the third image to obtain a first feature, and obtaining a first scoring result based on the first feature;

[0060] Step S500, extracting a second feature based on the first scoring result and the first feature, and fusing the second feature to obtain a fused feature;

[0061] Step S600, extracting features based on the fused features to obtain a third feature, and obtaining a second scoring result based on the third feature;

[0062] Step S700, extracting multi-scale features of the training face image, and obtaining a third scoring result according to the third feature, the second scoring result and the multi-scale features;

[0063] Step S800, adjusting the parameters of the face scoring model according to the first scoring result, the second scoring result and the third scoring result to obtain a trained face scoring model.

[0064] In step S200 of some embodiments, a face image to be trained is obtained. On the one hand, the face image can be taken by a photographic device such as a camera or a webcam, etc.; on the other hand, the face image can be obtained from a large online database.

[0065] By manually scoring the beauty of some face images, marking the scores on the face images, and obtaining the score labels of the face images, a first image marked with the score label and a second image not marked with the score label are obtained, and the first image and the second image constitute a set of face images to be trained.

[0066] The training face images are preprocessed by alignment, data augmentation, and standardization. For example, the training face images are preprocessed by data augmentation techniques such as rotation, scaling, translation, flipping, and color transformation to increase the diversity and robustness of the data.

[0067] The preprocessed face images to be trained are input into the face scoring model.

[0068] In step S300 of some embodiments, a pseudo label is generated according to the first image, and the pseudo label is assigned to the second image to obtain a third image annotated with the pseudo label.

[0069] Exemplarily, a pre-trained convolutional neural network, such as ResNet-50, is used to extract a multi-scale feature map of the face image to be trained, and the multi-scale feature map includes feature maps of Conv3, Conv4, and Conv5 layers.

[0070] Reference Figure 3 , generating a pseudo label according to the first image, including the following steps: Step S310, weakly supervised learning is performed on the classifier of the face scoring model according to the first image to generate a pseudo label. Specifically, the classifier of the face scoring model is trained using the first image annotated with the scoring label. The trained classifier is used to predict the second image not annotated with the scoring label to generate a pseudo label. Through weakly supervised learning, the image annotation cost can be reduced, and a large number of annotated images can be quickly obtained.

[0071] Reference Figure 4, assigning a pseudo label to the second image to obtain a third image annotated with the pseudo label, comprising the following steps:

[0072] Step S320, filtering the pseudo labels according to the confidences corresponding to the pseudo labels to obtain filtered pseudo labels;

[0073] Step S330: assigning the filtered pseudo-label to the second image to obtain a third image annotated with the pseudo-label.

[0074] Specifically, a first image labeled with a rating label is used to train a classifier of a face rating model. The trained classifier is used to predict a second image that is not labeled with a rating label to generate multiple pseudo labels. When generating pseudo labels, the confidence probability of each pseudo label is also generated. The pseudo labels are screened according to the confidence corresponding to the pseudo labels, and the pseudo labels with the highest confidence are generally selected. Each second image that is not labeled with a rating label is assigned a class with the highest probability as its pseudo label. This process can be enhanced by the entropy minimization method to encourage the model to make low-entropy predictions for unlabeled data. The lower the entropy of the prediction result, the more accurate it is.

[0075] Reference Figure 5 In step S400 of some embodiments, extracting features from the first image and the third image to obtain a first feature, and obtaining a first scoring result based on the first feature include the following steps:

[0076] Step S410, performing semi-supervised learning on a classifier of a face scoring model according to the first image and the third image, and extracting a first feature;

[0077] Step S420, performing classification prediction according to the first feature to obtain a first scoring result.

[0078] The generated pseudo labels are used together with the true labels to further train the model. In each iteration, select pseudo-label data with high confidence to add to the training set, remove prediction results with low confidence, and update the model. Evaluate the performance of the model trained with pseudo labels. In some methods, the weights of the pseudo-label part can be adjusted through a deterministic simulation process to avoid falling into a local optimum. The above process is iterated until the model performance no longer improves or the predetermined number of iterations is reached. In each iteration, the model may need to be fine-tuned to ensure the quality of the pseudo-labels and the generalization ability of the model.

[0079] The trained classifier performs feature extraction based on the first image and the third image to obtain a first feature, and performs classification prediction based on the first feature to obtain a first scoring result.

[0080] Reference Figure 6In step S500 of some embodiments, feature extraction is performed based on the first scoring result and the first feature to obtain a second feature, and feature fusion is performed based on the second feature to obtain a fusion feature, including the following steps:

[0081] Step S510, performing convolution on the first scoring result and the first feature to extract a second feature;

[0082] Step S520, converting the second feature to a feature space of the same dimension;

[0083] Step S530, performing feature fusion on the converted second features to obtain fused features.

[0084] Specifically, a convolutional neural network is used to perform convolution on the first scoring result and the first feature to extract the second feature. Multiple second features are in different dimensions; in order to unify the dimensions of the second features, multiple second features in different dimensions are converted to feature spaces of the same dimension through an embedding layer. The converted second features are subjected to feature fusion through a graph neural network (GNN) to obtain fused features.

[0085] It can be understood that graph neural network refers to a neural network that learns graph structured data and extracts and discovers features and patterns in graph structured data.

[0086] Reference Figure 7 In step S600 of some embodiments, feature extraction is performed based on the fusion feature to obtain a third feature, and a second scoring result is obtained based on the third feature, including the following steps:

[0087] Step S610, training a classifier of a face scoring model according to the fused features, mapping the fused features to a class activation map, and obtaining a third feature;

[0088] Step S620: Perform global maximum pooling according to the third feature to obtain a second scoring result.

[0089] Specifically, the classifier of the face scoring model is trained based on the fusion features, and the multi-scale fusion features are mapped to the class activation map through a convolution layer with a 1×1 convolution kernel to obtain the third feature. The third feature is subjected to global maximum pooling to extract the feature score of each category, which is the second scoring result.

[0090] The use of class activation maps and global maximum pooling enhances the interpretability of the model, which can clearly show the model's attention to different areas of the image and improve the credibility and visualization of the scoring.

[0091] Among them, Class Activation Mapping (CAM) is a technique used in deep learning to visualize the decision-making process of convolutional neural networks. It identifies the most critical areas in the image for classification decisions and is used to understand whether the model makes classification decisions based on visual or time series features. The generation of class activation maps is based on an observation that global average pooling simplifies the data but retains spatial or temporal information, which makes it possible to reversely infer the area that contributes most to the classification result through the feature map of the convolution layer and the weight of the corresponding category.

[0092] Specifically, the fused features are reduced to feature vectors through global average pooling; the feature vectors are multiplied by the weights of the fully connected layer and converted into probabilities through the softmax function; for a specific category, the class activation map is calculated by multiplying the weights of the corresponding category from the global average pooling layer to the softmax layer with the activation feature map of the last convolutional layer.

[0093] By fusing the multi-scale feature map and the class activation map through step S500 and step S600, the prediction accuracy is improved, and the prediction accuracy and the generalization ability of the model are significantly improved.

[0094] In step S700 of some embodiments, multi-scale features of the training face image are extracted, and a third scoring result is obtained based on the third feature, the second scoring result and the multi-scale features.

[0095] Specifically, the features of the training face images are obtained from different scales and levels to obtain multi-scale features, capturing various details including low-level edge features and high-level semantic features, ensuring that the model can fully understand the image content. The third feature, the second scoring result and the multi-scale feature are fused, and by combining the feature information and the scoring result, the complementarity between the features is used to improve the accuracy of the scoring and make more accurate judgments. After fusing multiple third features, the second scoring result and the multi-scale features, the third scoring result is generated through further calculation and analysis. The score integrates all feature information and optimization results to provide a comprehensive and reliable beauty evaluation.

[0096] Reference Figure 8 In step S800 of some embodiments, adjusting the parameters of the face scoring model according to the first scoring result, the second scoring result and the third scoring result to obtain a trained face scoring model includes the following steps:

[0097] Step S810, obtaining a target scoring result according to the first preset weight parameter, the second preset weight parameter, the third preset weight parameter, the first scoring result, the second scoring result and the third scoring result;

[0098] Step S820, obtaining a loss function value according to the target scoring result;

[0099] Step S830, adjusting the parameters of the face scoring model according to the loss function value to obtain a trained face scoring model.

[0100] Specifically, through weight allocation, different weights are allocated according to the importance and reliability of each feature to ensure that the contribution of each feature in the final score is reasonable; the first preset weight parameter, the second preset weight parameter, and the third preset weight parameter are configured for the first scoring result, the second scoring result, and the third scoring result, respectively. The first preset weight parameter is expressed as ω1, the second preset weight parameter is expressed as ω2, and the third preset weight parameter is expressed as ω3. And the first preset weight parameter, the second preset weight parameter, and the third preset weight parameter satisfy the following relationship: w1+w2+w3=1. The target scoring result can be expressed as: S=ω1*S1+ω2*S2+ω3*S3. Among them, S1 is the first scoring result, S2 is the second scoring result, and S3 is the third scoring result.

[0101] By integrating multiple scoring results for weighted fusion and combining the importance and reliability of different features, the accuracy and reliability of the target scoring results are ensured.

[0102] An embodiment of the present application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned face scoring model training method and face scoring method when executing the computer program.

[0103] The electronic device may be any intelligent terminal including a computer.

[0104] In general, for the hardware structure of the electronic device, the processor can be implemented using a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., to execute related programs to implement the technical solutions provided in the embodiments of the present application.

[0105] The memory can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory can store an operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory, and the processor is called to execute the methods of the embodiments of this application.

[0106] The input / output interface is used to realize information input and output.

[0107] The communication interface is used to realize the communication interaction between this device and other devices. Communication can be achieved through wired methods (such as USB, network cable, etc.) or wireless methods (such as mobile network, WIFI, Bluetooth, etc.).

[0108] The bus transmits information between the various components of the device (such as the processor, memory, input / output interface, and communication interface). The processor, memory, input / output interface, and communication interface communicate with each other within the device through the bus.

[0109] An embodiment of the present application provides a computer storage medium storing computer executable instructions for executing the above-mentioned face scoring model training method and face scoring method.

[0110] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium. In the above description of this specification, the description of reference terms "one embodiment / example", "another embodiment / example" or "certain embodiments / examples" and the like means that the specific features, structures, materials or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0111] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0112] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0113] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0114] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0115] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above-mentioned units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. Although the embodiments of the present application have been shown and described, it can be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and purposes of the present application, and the scope of the present application is defined by the claims and their equivalents.

[0116] The above is a specific description of the preferred implementation of the present application, but the present application is not limited to the embodiments. Technical personnel familiar with the field can make various equivalent modifications or substitutions without violating the spirit of the present application. These equivalent modifications or substitutions are all included in the scope defined by the claims of the present application.

Claims

1. A training method for a face scoring model, characterized in that, comprising: Obtain the face images to be trained, and input the face images to be trained into the face scoring model. The face images to be trained include a first image labeled with a scoring label and a second image without a labeled scoring label; Generate pseudo-labels according to the first image, and assign the pseudo-labels to the second image to obtain a third image labeled with pseudo-labels; Extract first features according to the first image and the third image, and obtain a first scoring result according to the first features; Extract second features according to the first scoring result and the first features, and perform feature fusion according to the second features to obtain fused features; Extract third features according to the fused features, and obtain a second scoring result according to the third features; Extract multi-scale features of the training face images, and obtain a third scoring result according to the third features, the second scoring result, and the multi-scale features; Adjust the parameters of the face scoring model according to the first scoring result, the second scoring result, and the third scoring result to obtain a trained face scoring model.

2. The training method for a face scoring model according to claim 1, characterized in that, The generating pseudo-labels according to the first image includes: Performing weak supervision learning on the classifier of the face scoring model according to the first image to generate pseudo-labels.

3. The training method for a face scoring model according to claim 1, characterized in that, The assigning the pseudo-labels to the second image to obtain a third image labeled with pseudo-labels includes: Screen the pseudo-labels according to the confidence corresponding to the pseudo-labels to obtain the screened pseudo-labels; Assign the screened pseudo-labels to the second image to obtain a third image labeled with pseudo-labels.

4. The training method for a face scoring model according to claim 1, characterized in that, The extracting first features according to the first image and the third image, and obtaining a first scoring result according to the first features includes: Performing semi-supervised learning on the classifier of the face scoring model according to the first image and the third image to extract first features; Performing classification prediction according to the first features to obtain a first scoring result.

5. The training method for a face scoring model according to claim 1, characterized in that, The extracting second features according to the first scoring result and the first features, and performing feature fusion according to the second features includes: Performing convolution according to the first scoring result and the first features to extract second features; Converting the second features to a feature space of the same dimension; Performing feature fusion on the converted second features to obtain fused features.

6. The training method for a face scoring model according to claim 1, characterized in that, The extracting third features according to the fused features, and obtaining a second scoring result according to the third features includes: Train a classifier of the face scoring model according to the fusion features, map the fusion features to a class activation map, and obtain a third feature; Perform global max pooling according to the third feature to obtain a second scoring result.

7. The training method of the face scoring model according to claim 1, characterized in that the adjusting the parameters of the face scoring model according to the first scoring result, the second scoring result and the third scoring result to obtain a trained face scoring model includes: obtaining a target scoring result according to a first preset weight parameter, a second preset weight parameter, a third preset weight parameter, the first scoring result, the second scoring result and the third scoring result; obtaining a loss function value according to the target scoring result; adjusting the parameters of the face scoring model according to the loss function value to obtain a trained face scoring model.

8. A face scoring method, characterized in that it includes: obtaining a face image to be scored; inputting the face image to be scored into a trained face scoring model for scoring to obtain a scoring result; wherein, the trained face scoring model is trained according to the training method of the face scoring model according to any one of claims 1 to 7.

9. An electronic device, characterized in that the electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the training method of the face scoring model according to any one of claims 1 to 7 and the face scoring method according to claim 8 are implemented.

10. A computer-readable storage medium stores a computer program, characterized in that when the computer program is executed by a processor, the training method of the face scoring model according to any one of claims 1 to 7 and the face scoring method according to claim 8 are implemented.

Images